Developing AI Integration in Domain Specific Tools with Theia AI
summary
Demo shows Theia AI framework embedded in Eclipse Theia wiring LLMs, prompt templates, function calls and context augmentation so IDE features are reusable. A GLSP diagram editor plus chat answers “max parallel tasks” correctly, creates nodes, and updates both diagram and underlying documentation while showing a diff — the demo works end-to-end using an OpenAI model (GPT‑4). Architecture supports agents and variable providers, dynamic RAG-style context selection, and tool calling and RAG with options for front-end, back-end, or on‑prem LLMs.
AI-written summary · about
chapters (16)
- 0:00AI integration in domain tools
- 1:03AI capabilities in custom tools
- 4:56Theia AI framework overview
- 6:21Agents as tool-LLM mediators
- 7:05Prompt engineering for tools
- 9:15Creating agents in Theia AI
- 10:44Context in agent prompts
- 11:48Dynamic context with variables
- 13:33Retrieval augmented generation
- 14:58Multi-term prompting
- 16:23Classification prompts in code
- 18:53Actionable LLM responses
- 20:20Tool calling workflows
- 21:44AI beyond chat agents
- 22:45AI integration recommendations
- 23:47Q&A
related talks
15mHow Theia AI is Redefining the Developer Workflow
36mApplied AI with Eclipse Graphene
58mAI Assistant: Don’t Code Alone
25mVirtual Tool Calling and the Future of Portable AI Toolchains
01h 23mBuild your own AI Agent
26mThe Developer’s New Teammate: AI in Embedded and IoT Environments
14mDeveloper productivity in a post-AI world
46mAI Integration for Devs: ChatGPT and Copilot in Your Projects